USE OF ARTIFICIAL INTELLIGENCE IN CUSTOMS PROCEDURES

Recently, the use of artificial intelligence (AI) in customs procedures has become a trending topic particularly the use of AI for determining tariff classifications of goods, and even for filling out customs declarations. While we are still at the early stages of implementing AI in this field, there is widespread agreement that the future holds developments beyond our imagination. Of course, no one can predict the future, but under current conditions, it's worth discussing the role of AI in customs processes. It may even be time though perhaps still early to ask: "Will AI make customs brokerage obsolete?"

Before diving into such a discussion, it's essential to define the concepts involved. In order to have a meaningful debate, we need to agree on what terms like “intelligence,” “reason,” and “artificial intelligence” actually mean. Establishing a common understanding will ensure we’re speaking the same language.

Reason vs. Intelligence

In philosophical terms, reason is the capacity to form concepts and make judgments based on them in other words, to understand and evaluate a situation holistically. It is the ability to discern meaning and to distinguish between right and wrong, truth and falsehood. Reason tends to develop with age and maturity.

Intelligence, on the other hand, is the mind’s ability to learn, apply what’s learned, adapt to new situations, and devise new solutions. It encompasses multiple mental functions such as perception, memory, reasoning, and learning.

The key difference? Intelligence can be measured (such as with IQ tests), while reason is more abstract and ethical. Intelligence reflects how quickly the brain can process information, but it doesn't necessarily imply ethical decision-making. A person may be highly intelligent but not always make the right choices. A composer may use their intelligence to create complex music but struggle with basic math problems. Intelligence allows for problem-solving; reason governs the application of those solutions in meaningful ways.

Artificial Intelligence

Artificial intelligence is a type of intelligence executed by machines, as opposed to natural (human) intelligence. AI refers to computer systems or software that can mimic human cognitive functions — learning, problem-solving, understanding language, speech, visual recognition, and more.

Machines surpass humans in data storage and memory retention. However, reasoning, perception, and problem-solving still largely require the human brain. Therefore, AI can be described as a combination of information storage capacity and the ability to imitate human intelligence. For AI to work effectively, machines must be equipped with the ability to learn. Thus, AI systems can:

· Learn from data (machine learning),

· Apply logic,

· Make decisions,

· Understand and process natural language,

· Recognize images.

Given that we are only at the beginning of this technological revolution, we cannot predict how far AI will develop or how it will evolve. This uncertainty is what raises concerns about the potentially dangerous outcomes of AI.

Use of AI in Customs Procedures

In customs, AI is increasingly used to reduce workload and errors, speed up processes, and enhance security. Specific applications include:

· Completing customs declarations,

· Reading and digitizing documents such as invoices, bills of lading, and certificates of origin,

· Determining the Customs Tariff Statistical Position (HS Code) of goods,

· Checking draft declarations to minimize penalty risks,

· Translating documents and their contents.

What Are the Risks?

OCR (Optical Character Recognition) technology converts printed documents into digital images and identifies each character. Initially limited to word recognition, today’s OCR systems can read a variety of languages, formats, and even handwritten texts. These texts are then processed with Natural Language Processing (NLP), allowing AI to analyze not just words but also context and meaning to categorize and process them correctly.

However, AI's interpretation of these documents may not always align with human understanding. Unverified digital data can lead to critical errors.

Machine translation is still not perfect. Especially with synonyms or idiomatic expressions, translation errors are common.

Determining tariff codes based solely on product names can be misleading. Often, the composition and intended use of goods are critical to classification. Relying exclusively on AI for tariff determination may result in severe penalties.

In Summary

AI can significantly reduce workload and accelerate customs procedures. However, over-reliance on AI carries substantial risks, particularly in legal and regulatory contexts.

The Dangers

As noted earlier, AI learns continuously. This means that every document and piece of information introduced into the system becomes accessible — at least indirectly — to the software developer or potentially other users. This raises serious concerns about data privacy.

Additionally, AI reduces the human workload by performing tasks traditionally done by people. For instance, in the film industry, screenwriters protested in Hollywood, claiming AI was taking their jobs. This scenario is applicable across many sectors, including customs.